Do clustering monoclonal antibody solutions really have a concentration dependence of viscosity?

Abstract

Protein solution rheology data in the biophysics literature have incompletely identified factors that govern hydrodynamics. Whereas spontaneous protein adsorption at the air/water (A/W) interface increases the apparent viscosity of surfactant-free globular protein solutions, it is demonstrated here that irreversible clusters also increase system viscosity in the zero shear limit. Solution rheology measured with double gap geometry in a stress-controlled rheometer on a surfactant-free Immunoglobulin solution demonstrated that both irreversible clusters and the A/W interface increased the apparent low shear rate viscosity. Interfacial shear rheology data showed that the A/W interface yields, i.e., shows solid-like behavior. The A/W interface contribution was smaller, yet nonnegligible, in double gap compared to cone-plate geometry. Apparent nonmonotonic composition dependence of viscosity at low shear rates due to irreversible (nonequilibrium) clusters was resolved by filtration to recover a monotonically increasing viscosity-concentration curve, as expected. Although smaller equilibrium clusters also existed, their size and effective volume fraction were unaffected by filtration, rendering their contribution to viscosity invariant. Surfactant-free antibody systems containing clusters have complex hydrodynamic response, reflecting distinct bulk and interface-adsorbed protein as well as irreversible cluster contributions. Literature models for solution viscosity lack the appropriate physics to describe the bulk shear viscosity of unstable surfactant-free antibody solutions.

Figure 1

Schematic of preferential adsorption of surfactant at air/water interface (orogenic displacement). Surfactant molecules (red) have a hydrophobic head and hydrophilic tail, and monoclonal antibody molecules (green) have a Y-shape (Fab and Fc domains).

Figure 2

(a) Cone-plate rheometry data of as-is solutions (taken from the storage refrigerator and allowed to equilibrate to 23°C in the rheometer). Data are shown on the stock solution (107 g L⁻¹) and diluted (10, 26, 50, and 70 g L⁻¹) solutions. (Inset) Viscosity values at 10.4 s⁻¹ (circles) and 1040 s⁻¹ (squares) as a function of protein concentration. (Solid and open symbols) Unfiltered and filtered solutions, respectively. (b) Flow curves of unfiltered solution (solid symbols) and filtered solution (open symbols) at 23°C for 10 g L⁻¹ and 107 g L⁻¹ mAb solutions. (Inset plot) Shear stress versus shear rate.

Figure 3

Flow curves at 23°C of unfiltered mAb solution (107 g L⁻¹) in the cone-plate geometry (blue squares), unfiltered solution in the double gap (red squares), and filtered solution in the double gap geometry (green squares). (Asterisks; same color code applies) Shear stress.

Figure 4

Effects of filtration on the flow curves of surfactant-spiked antibody solution ([mAb] = 100 g L⁻¹) at 5°C. (Squares and diamonds) Viscosity and shear stress, respectively. (Blue and red symbols) Unfiltered and filtered solutions, respectively.

Figure 5

HP-SEC peaks of filtered and unfiltered solutions showing the presence of peaks due to the monomer and reversible clusters. Monomer content is 98% in both cases. Data are only shown on the unfiltered solution for clarity and space. Filtered solutions gave quantitatively identical HP-SEC data as unfiltered solutions.

Figure 6

Autocorrelation function of intensity fluctuations (from dynamic light scattering) of filtered and unfiltered antibody solutions. (Symbols) Experimental data. (Solid lines) Exponential fits (Eq. 2).

Figure 7

Cluster size distribution (MFI data) of (a) unfiltered surfactant-free, (b) filtered surfactant-free, and (c) unfiltered surfactant-laden antibody solutions with representative optical micrographs (inset). Scale bar for micrograph in panel a also applies to panels b and c.

Figure 8

Peclèt number calculated from MFI data for unfiltered and filtered solutions, plotted versus shear rate.

Figure 9

Surface flow curve (surface viscosity and surface shear stress) of mAb solution, compared to BSA data of Sharma et al. (17).